Skip to contentSkip to conversation

New: intel_pstate does not obey min/max frequencies set by user

Bug 86461 [email protected] 1 participant 8 messages Oct 17, 2014 – Oct 28, 2014

lore
Share & tools
lore

[email protected]

author
#0 lore raw

https://bugzilla.kernel.org/show_bug.cgi?id=86461

            Bug ID: 86461
           Summary: intel_pstate does not obey min/max frequencies set by
                    user
           Product: Power Management
           Version: 2.5
    Kernel Version: 3.17-rc5
          Hardware: All
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: normal
          Priority: P1
         Component: cpufreq
          Assignee: [email protected]
          Reporter: [email protected]
        Regression: No
# uname -a    
Linux aep-vaio 3.17.0-rc5 #1 SMP PREEMPT Thu Sep 18 23:13:08 YEKT 2014 x86_64
Intel(R) Core(TM) i7-2640M CPU @ 2.80GHz GenuineIntel GNU/Linu
# echo 2400000 | tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_min_freq
# cpufreq-info
cpufrequtils 008: cpufreq-info (C) Dominik Brodowski 2004-2009
Report errors and bugs to [email protected], please.
analyzing CPU 0:
  driver: intel_pstate
  CPUs which run at the same hardware frequency: 0
  CPUs which need to have their frequency coordinated by software: 0
  maximum transition latency: 0.97 ms.
  hardware limits: 800 MHz - 3.50 GHz
  available cpufreq governors: performance, powersave
  current policy: frequency should be within 2.40 GHz and 3.50 GHz.
                  The governor "powersave" may decide which speed to use
                  within this range.
  current CPU frequency is 804 MHz (asserted by call to hardware).
analyzing CPU 1:
  driver: intel_pstate
  CPUs which run at the same hardware frequency: 1
  CPUs which need to have their frequency coordinated by software: 1
  maximum transition latency: 0.97 ms.
  hardware limits: 800 MHz - 3.50 GHz
  available cpufreq governors: performance, powersave
  current policy: frequency should be within 2.40 GHz and 3.50 GHz.
                  The governor "powersave" may decide which speed to use
                  within this range.
  current CPU frequency is 885 MHz (asserted by call to hardware).
analyzing CPU 2:
  driver: intel_pstate
  CPUs which run at the same hardware frequency: 2
  CPUs which need to have their frequency coordinated by software: 2
  maximum transition latency: 0.97 ms.
  hardware limits: 800 MHz - 3.50 GHz
  available cpufreq governors: performance, powersave
  current policy: frequency should be within 2.40 GHz and 3.50 GHz.
                  The governor "powersave" may decide which speed to use
                  within this range.
  current CPU frequency is 800 MHz (asserted by call to hardware).
analyzing CPU 3:
  driver: intel_pstate
  CPUs which run at the same hardware frequency: 3
  CPUs which need to have their frequency coordinated by software: 3
  maximum transition latency: 0.97 ms.
  hardware limits: 800 MHz - 3.50 GHz
  available cpufreq governors: performance, powersave
  current policy: frequency should be within 2.40 GHz and 3.50 GHz.
                  The governor "powersave" may decide which speed to use
                  within this range.
  current CPU frequency is 822 MHz (asserted by call to hardware).

As you see, my attempt to force the minimum frequency to 2.4 GHz has been ignored.

Signature · 2 lines
You are receiving this mail because:
You are the assignee for the bug.
author
#1 lore raw

[Bug 86461] intel_pstate does not obey min/max frequencies set by user

https://bugzilla.kernel.org/show_bug.cgi?id=86461

Dirk Brandewie <[email protected]> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |[email protected]

--- Comment #1 from Dirk Brandewie <[email protected]> --- Can you send the output of the following commands before and after setting scaling_min_freq:

turbostat sleep 5
grep MH /proc/cpuinfo
grep . /sys/devices/system/cpu/intel_pstate/*

Signature · 2 lines
You are receiving this mail because:
You are the assignee for the bug.
author
#2 lore raw

[Bug 86461] intel_pstate does not obey min/max frequencies set by user

In reply to [email protected]

https://bugzilla.kernel.org/show_bug.cgi?id=86461

--- Comment #2 from Alexander E. Patrakov <[email protected]> --- aep-vaio turbostat # ./turbostat sleep 5
Core CPU Avg_MHz %Busy Bzy_MHz TSC_MHz SMI CPU%c1 CPU%c3
CPU%c6 CPU%c7 CoreTmp PkgTmp Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 PkgWatt CorWatt GFXWatt
- - 32 2.47 1298 2794 0 5.02 0.04
0.00 92.47 50 50 3.79 0.02 4.49 77.75 4.43 0.74
0.45

       0       0      37    2.97    1253    2794       0    6.90    0.07   
0.00   90.06      50      50    3.79    0.02    4.49   77.75    4.43    0.74   
0.45
       0       1      33    2.11    1558    2794       0    7.77
       1       2      27    3.05     895    2794       0    2.06    0.00   
0.00   94.89      50
       1       3      31    1.76    1763    2794       0    3.35

5.004419 sec
aep-vaio turbostat # grep MH /proc/cpuinfo
cpu MHz : 951.781
cpu MHz : 835.625
cpu MHz : 852.468
cpu MHz : 822.500
aep-vaio turbostat # grep . /sys/devices/system/cpu/intel_pstate/* /sys/devices/system/cpu/intel_pstate/max_perf_pct:100 /sys/devices/system/cpu/intel_pstate/min_perf_pct:22 /sys/devices/system/cpu/intel_pstate/no_turbo:0
aep-vaio turbostat # echo 2400000 | tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_min_freq
2400000
aep-vaio turbostat # ./turbostat sleep 5
Core CPU Avg_MHz %Busy Bzy_MHz TSC_MHz SMI CPU%c1 CPU%c3
CPU%c6 CPU%c7 CoreTmp PkgTmp Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 PkgWatt CorWatt GFXWatt
- - 39 1.74 2229 2794 0 4.47 0.51
0.00 93.27 53 53 4.40 0.23 5.35 77.63 4.48 0.87
0.37

       0       0      48    2.42    1981    2794       0    2.50    0.61   
0.00   94.47      53      53    4.40    0.23    5.35   77.63    4.48    0.87   
0.37
       0       1      13    0.78    1727    2794       0    4.14
       1       2      45    1.84    2420    2794       0    5.66    0.42   
0.00   92.08      53
       1       3      50    1.93    2560    2794       0    5.57

5.001705 sec
aep-vaio turbostat # grep MH /proc/cpuinfo
cpu MHz : 1277.390
cpu MHz : 1711.718
cpu MHz : 1661.734
cpu MHz : 1932.875
aep-vaio turbostat # grep . /sys/devices/system/cpu/intel_pstate/* /sys/devices/system/cpu/intel_pstate/max_perf_pct:100 /sys/devices/system/cpu/intel_pstate/min_perf_pct:68 /sys/devices/system/cpu/intel_pstate/no_turbo:0

Signature · 2 lines
You are receiving this mail because:
You are the assignee for the bug.
author
#4 lore raw

[Bug 86461] intel_pstate does not obey min/max frequencies set by user

In reply to [email protected]

https://bugzilla.kernel.org/show_bug.cgi?id=86461

--- Comment #4 from Dirk Brandewie <[email protected]> --- I looks like everything is working correctly. You have to remember that the frequencies returned by turbostat and intel_pstate are measured effective frequency over the time of the sample. Previously with acpi_cpufreq the value returned through /proc/cpuinfo and the cupfreq sysfs interface was the requested
frequency and not the actual frequency.

If you want to see that the floor for P states has actually changed you can see the requested P state with:
rdmsr -a -f15:8 0x199

The actual frequency that all cores on the processor run at is selected by the processor based on the requests for all cores.

Signature · 2 lines
You are receiving this mail because:
You are the assignee for the bug.
author
#5 lore raw

[Bug 86461] intel_pstate does not obey min/max frequencies set by user

In reply to [email protected]

https://bugzilla.kernel.org/show_bug.cgi?id=86461

--- Comment #5 from Alexander E. Patrakov <[email protected]> --- OK, the floor for the P states has indeed changed. If that MSR value corresponds to the CPU frequency in units of 100 MHz, then the value is indeed the desired one, and the bug can be closed as invalid.

Signature · 2 lines
You are receiving this mail because:
You are the assignee for the bug.
author
#6 lore raw

[Bug 86461] intel_pstate does not obey min/max frequencies set by user

In reply to [email protected]

https://bugzilla.kernel.org/show_bug.cgi?id=86461

Alan <[email protected]> changed:

What |Removed |Added
----------------------------------------------------------------------------

             Status|NEW                         |RESOLVED
                 CC|                            |[email protected]
         Resolution|---                         |INVALID
Signature · 2 lines
You are receiving this mail because:
You are the assignee for the bug.